The location of the divalent metal binding sites and the light chain

DTNB light chain gives rise to an EPR spectrum similar to that of Mn(II) bound to myosin and this indicatesthat the metal binding site comprises ligan...
0 downloads 0 Views 3MB Size
MYOSIN

LIGHT CHAINS A N D METAL BINDING

On the Location of the Divalent Metal Binding Sites and the Light Chain Subunits of Vertebrate Myosin? Clive R. Bagshaw*

ABSTRACT:

The divalent metal ion binding sites of skeletal myosin were investigated by electron paramagnetic resonance (EPR) spectroscopy using the paramagnetic Mn(I1) ion as a probe. Myosin possesses two high affinity sites ( K < 1 pM) for Mn(lI), which are located on the 5,5’-dithiobis(2-nitrobenzoate) (DTNB) light chains. Mn(I1) bound to the isolated DTNB light chain gives rise to an EPR spectrum similar to that of Mn(I1) bound to myosin and this indicates that the metal binding site comprises ligands from the D T N B light chain alone. Myosin preparations in which the D T N B light chain content is reduced by treatment with 5,5’-dithiobis(2-nitrobenzoate) show a corresponding reduction in the stoichiometry of Mn(I1) binding, but the stoichiometry is recovered on reassociation of the DTNB light chain. Chymotryptic digestion of myosin filaments in the presence of ethylenediaminetetraacetic acid yields subfragment 1, but digestion in the presence of divalent metal ions produces heavy meromyosin. Myosin with a depleted D T N B light chain content gives rise to subfragment 1 on proteolysis, even in the presence of divalent metal ions. It is proposed that saturation of the DTNB light chain site with divalent ions protects this subunit against proteolysis, which, in turn, inhibits the cleavage of the subfragment 1-subfragment 2 link. Either the DTNB light chain

is located near the region of the link and sterically blocks chymotryptic attack, or it is bound to the subfragment 1 moiety and affects the conformation of the link region. When the product heavy meromyosin was examined by sodium dodecyl sulfate gel electrophoresis, an apparent anomaly arose in that there was no trace of the 19 000-dalton band corresponding to the DTNB light chain. This was resolved by following the time course of chymotryptic digestion of the myosin heavy chain, the DTNB light chain, and the divalent metal binding site. The 19 000-dalton DTNB light chain is rapidly degraded to a 17 000-dalton fragment which comigrates with the alkali 2 light chain. The divalent metal site remains intact, despite this degradation, and the 17 000 fragment continues to protect the subfragment 1-subfragment 2 link. In the absence of divalent metal ions, the 17 000-dalton fragment is further degraded and attack of the subfragment l link ensues. Mn(I1) bound to cardiac myosin gives an EPR spectrum basically similar to that of skeletal myosin, suggesting that their 19 000-dalton light chains are analogous with respect to their divalent metal binding sites, despite their chemical differences. The potential of EPR spectroscopy for characterizing the metal binding sites of myosin from different sources and of intact muscle fibers is discussed.

C o n t r a c t i o n of vertebrate skeletal muscle is initiated through the release of Ca(II), which binds to the troponintropomyosin complex associated with the actin filament, thereby relieving the inhibition of the actin-activated myosin ATPase activity (see Weber and Murray, 1973). The existence of Ca(I1) binding sites on the myosin thick filaments has raised the possibility that other events are controlled by Ca(I1) release (Morimoto and Harrington, 1974; Bremel and Weber, 1975). However, as yet there is no clear evidence that vertebrate skeletal muscle possesses a myosin-linked regulatory system of the type found in many invertebrate muscles (Lehman and Szent-Gyorgyi, 1975) since, in the absence of the troponintropomyosin components, Ca(I1) has no major effect on the in vitro actomyosin ATPase rate. As a basis for discussion of the significance of the Ca(I1) binding sites of myosin, it is pertinent to review the structural features of this molecule. Myosin (480 000 daltons) is a hexamer comprising two heavy chains (200 000 daltons) and four light chains (approximately 20 000 daltons). Each heavy chain folds to make up a roughly globular region and a long filamentous region which intercoils with its partner heavy chain to form a rod. At

physiological ionic strength, the rods aggregate to provide the backbone of the myosin thick filament. Two light chains are associated with each globular region of the myosin molecule. This structure emerged from studies in which the myosin molecule was subjected to limited proteolytic digestion (Lowey et al., 1969). A susceptible region is present in the myosin rod which, on proteolysis, yields light meromyosin (140 000 daltons) and a water-soluble fragment, heavy meromyosin (340 000 daltons). Heavy meromyosin contains the two globular moieties (subfragment I , 2 X 1 10 000 daltons) connected by a short rod region (subfragment 2,60 000 daltons) and may be further digested to release these fragments. Each subfragment 1 bears one ATPase site and one to two light chains. Since these products can be obtained by digestion with a number of proteolytic enzymes of differing specificity, it is thought that the peptide chain may be uncoiled in the susceptible regions and, hence, these may represent the positions of flexible joints in the myosin filament (Huxley, 1969). In the case of vertebrate fast skeletal muscle, three types of light chains have been identified (21 000, 19 000,and 17 000 daltons), giving rise to the possibility of myosin isoenzymes. The 19 000-dalton component can be partially but selectively removed from skeletal myosin, without loss of ATPase activity, by a combined treatment with DTNB’ and EDTA and is

+ From the Department of Biochemistry and Biophysics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania I9 174. Receiced June 29, 1976. This work was done during the tenure of a Research Fellowship of the Muscular Dystrophy Associations of America and wassupported by a National lnstitutesof Health Grant HL15692 and a travel grant from the Wellcome Trust, United Kingdom. * Address correspondence to: Department of Biochemistry, University of Oxford, Oxford OX1 3QU, U. K.



Abbreviations used: AI-LC, alkali I light chain; A2-LC. alkali 2 light chain: DTNB. 5.5’-dithiobis(2-nitrobenzoate):HMM-HC. heavv meromyosin heavy chain; PhCH2SOzF. phenylmethancsulfonyl fluoride; SI-HC, subfragment-1 heavy chain: EPR, electron paramagnetic resonance, EDTA ethylenediaminetetraacetic acid. BIOCHEMISTRY. VOL.

16, N O . 1, 1977

59

termed the D T N B light chain (Gazith et al., 1970; Weeds and Lowey, 197 1). Myosin contains two D T N B light chains and binds 2 mol of Ca(l1) with high affinity. There is growing evidence that the high affinity Ca(I1) sites of myosin are associated with the DTNB light chains (Werber et al., 1972, 1973; Morimoto and Harrington, 1974; Margossian et al., 1975). Werber et al. (1972) speculated that these subunits are located near each subfragment 1-subfragment 2 joint because the DTNB light chain is degraded during the papain digestion of myosin to yield subfragment 1, but it is retained in heavy meromyosin prepared by tryptic digestion. However, Margossian et al. (1975) showed that subfragment 1 could be prepared with a nearly stoichiometric DTNB light chain content, provided divalent metal ions were present during the papain digestion. In the present study the relationships between divalent metal ion binding, the D T N B light chain, and the susceptibility of the subfragment 1 -subfragment 2 joint to proteolysis are reexamined taking advantage of two recent experimental findings. During an investigation of metal binding to subfragment I , Bagshaw and Reed (1976) found a high affinity site (termed L site) for the paramagnetic Mn(I1) ion which was different from the active site. The stoichiometry of the L site depended on the conditions employed for the papain digestion of myosin and was found to correlate with the DTNB light chain content. Beinfeld et al. (1975) demonstrated that Ca(I1) was more effective than Mg(l1) in displacing Mn(I1) from the high affinity site. Hence Mn(I1) appears to be an analogue of Ca(1l) in binding to the D T N B light chain site (Bagshaw and Reed, 1976). EPR spectroscopy being a reasonably rapid and sensitive technique (-1 min per scan and -1 1 M free Mn(H20)6'+) offers a convenient method to study this site. In this particular case EPR spectroscopy presents a further advantage in that the X-band signal of Mn(I1) bound to the myosin L site is not extensively broadened, as is frequently observed for macromolecular complexes of Mn( 11) (Cohn and Townsend, 1954; Reed and Ray, 197 1). Thus under appropriate conditions EPR spectroscopy can be used to monitor, continuously and specifically, the occupancy and structure of this site with Mn(l1). Recently Yagi and Otani (1974) have established conditions for the proteolysis of myosin, by chymotrypsin, which do not lead to extensive cleavage of the heavy chain as is observed with trypsin (BBlint et al., 1975) or papain (Stone and Perry, 1973). While these latter enzymes yield apparently homogeneous heavy meromyosin and subfragment 1 preparations, the polypeptide chain is nicked, so that under the dissociating conditions of sodium dodecyl sulfate gel electrophoresis, a multi-banded heavy chain pattern is observed which complicates the identification of the digestion products. An interesting property of chymotryptic attack was revealed when myosin filaments were digested. Subfragment 1 was produced when the digestion was carried out in the presence of EDTA, but heavy meromyosin was the product in the presence of divalent cations (Weeds and Taylor, 1975). Other studies have demonstrated an effect of divalent metal ions on the digestion of myosin by trypsin, but the site of metal binding was not identified (BirB et al., 1972; Bhlint et al., 1975). A plausible mechanism for these findings is that saturation of the divalent metal site protects the D T N B light chain from digestion and, in turn, this subunit blocks proteolytic attack of the subfragment I-subfragment 2 joint. This hypothesis is tested since i t may clarify the mode of interaction between divalent metal ions and the myosin thick filament. The light chain subunits of myosin extracted from other

60

BIOCHEMISTRY. VOL.

16,

NO

I, 1977

muscles differ from those of fast skeletal muscle. Typically a component of about 19 000 daltons is present, although it may not be susceptible to dissociation by DTNB treatment. While differences in the I9 000-dalton components are apparent from their precise electrophoretic mobilities on sodium dodecyl sulfate gels and their peptide maps (Lowey and Risby, 197 1 : Weeds and Pope, 197 I ) , there are also analogies between them in their ability to be phosphorylated by a specific kinase (Frearson and Perry, 1975) and their ability to hybridize with desensitized invertebrate myosins and restore the Ca( 11) sensitivity of the myosin-linked regulatory system (Kendrick-Jones et al., 1976). It is of interest to investigate the EPR spectrum of Mn(II) bound to myosin isolated from other sources. This may provide a rapid and sensitive probe for examining the divalent metal sites of light chains analogous to the D T N B light chain of fast skeletal muscle. A preliminary study with rabbit cardiac myosin is reported. Materials and Methods Proteins. Skeletal myosin was prepared from the back and leg muscles of the rabbit by extraction of a muscle mince with 0.3 M KCI-0.15 M potassium phosphate and 1 mM ATP and purified by repeated dilution-precipitation steps essentially as described by Szent-Gyorgyi ( 1947). Cardiac myosin was prepared similarly from rabbit hearts. DTNB-EDTA treatment of myosin was carried out as described by Weeds and Lowey ( 1 971). After the 10-min reaction time, the myosin solution was either diluted tenfold to lower the ionic strength and precipitate the myosin with a reduced DTNB light chain content (DTNB-EDTA myosin), or it was diluted tenfold with 0.5 M NaCl and treated with 2 mM dithiothreitol to allow the DTNB light chain to reassociate (reconstituted myosin). As a control, myosin was treated with DTNB in the presence of 10 m M C a ( l l ) and was isolated by a dilution-precipitation step before treatment with dithiothreitol (DTNB-Ca( 11) myosin). After these treatments, all the myosin samples were dialyzed against 0.5 M NaC1-I inM dithiothreitol-50 mM Tris adjusted to pH 8.4 with HCI a t 4 "C for 5 h to unblock the thiol groups and then dialyzed against 20 mM Tris-HCI at pH 8.4. The myosin precipitates were collected by centrifugation and then washed three times with 20 m M Tris-HCI to remove unbound light chains and the residual DTNB. Finally the myosin precipitates were dissolved in 0.5 M NaCl to a concentration of about 15 mg/ml and dialyzed against 0.5 M NaCI-50 m M Tris-HCI at pH 8.4. Myosin concentrations were determined by the method of Lowry et al. ( I 95 1) which had been previously standardized by micro-Kjeldahl analysis. The supernatant from the DTN B-EDTA treatment, which contained the DTNB light chains, was treated with 4 m M dithiothreitol and was concentrated in a Minicon B15 macrosolute ultrafilter. it was washed free of 5-thio-2-nitrobenzoate in this apparatus with 1 m M dithiothreitol-50 m M Tris-HCI at pH 8.4. Chymotryptic Digestion. Myosin filaments were formed by dialysis against 0. I2 M NaCI-20 mM sodium phosphate a t pH 7.0 and were digested with a-chymotrypsin (type I-s from bovine pancreas, Sigma Chemical Co.) as described by Weeds and Taylor (1975). Digestion was terminated by addition of PhCH2SOlF; nevertheless, some residual proteolytic activity was noted which caused further degradation over a period of several days. The soluble products of digestion were recovered by dialysis of the digest against 5 m M sodium phosphate (pH 7.0) and removal of the insoluble protein by centrifugation.

MYOSIN

LIGHT CHAIYS A N D METAL

BINDING

Sodium Dodecyl Sulfate Gel Electrophoresis. Samples were prepared for sodium dodecyl sulfate gel electrophoresis by mixing a solution of the protein (1-4 mg/ml) with a n equal volume of a solution containing 2% sodium dodecyl sulfate, 2% 0-mercaptoethanol, 20% glycerol, 20 m M sodium phosphate at p H 7.0, and bromophenol blue tracking dye. The mixture was immediately heated to 100 O C for 2 min to circumvent residual proteolytic activity. Discontinuous zonal electrophoresis was carried out on 7.5 or 12.5% acrylamide gels in Tris-glycine buffer and in the presence of sodium dodecyl sulfate, essentially as described by Potter (1 974). The gels were stained with 0.1% fast green and scanned in a Gilford 240 spectrophotometer with a linear transport attachment. The molecular weights of the myosin components were calculated from their mobilities using phosphorylase a , actin, and hemoglobin as standards (Weber and Osborn, 1969). The values reported here serve as a means of identifying components rather than providing an accurate measure of their molecular weights. E P R Spectroscopy. EPR spectra were recorded at 9.0 G H z (X-band) using a Varian E-3 spectrometer. Samples (approximately 50 pl) were contained in high purity quartz capillary tubing. The six-line EPR spectrum characteristic of Mn(H20)62+ in aqueous solution generally becomes extensively broadened on complex formation and allows binding to be monitored in a titration simply by following the disappearance of the free Mn(I1) signal (Cohn and Townsend, 1954). However, Mn(I1) bound to the high affinity site (L site) of myosin contributes significantly to the observed EPR signal (Yazawa and Morita, 1974; Bagshaw and Reed, 1976). At 20 O C the amplitude of the signal of Mn(I1) bound to the L site is approximately 20% of that of an equivalent concentration of free Mn(H20)6’+and this percentage is increased a t 0 O C because only the free Mn(Hz0)62+ signal is reduced by the temperature change. While the L-site spectrum is not extensively broadened, it differs from that of Mn(H20)62+ in that the amplitudes of the inner peaks are reduced relative to the outer peaks (see, for example, Figure 8). The relative contribution from Mn(Hz0)62+ and Mn(I1) bound a t the L site therefore may be evaluated using eq 1 Robsd - RL (1) Rf - RL where Plf is the fraction of the observed amplitude of the 1st (low-field) peak arising from the free M n ( H ~ 0 ) 6 ~R+o,b s d is the observed ratio of the amplitudes of the 2nd to 1st peak, R L is the ratio for the L-site alone, and Rf is the ratio for free Mn(H20)62+.Typically R L = 0.6 and Rf = 1.1, but these ratios were checked for each experiment since the precise values showed some dependence (-5% variation) on the spectrometer adjustments. R L was obtained from the observed spectrum of a Mn(I1)-myosin mixture which had been equilibrated with a known low concentration of free Mn(H20)6z+ by dialysis. Titrations were carried out by adding a known concentration of MnC12 (30- 120 pM) to solutions of myosin (1 5-20 mg/ml) and recording the EPR spectra. The amount of myosin added (-100 p l ) was checked by weight to avoid the pipetting error incurred with such viscous solutions. The spectra were analyzed according to eq 1. Mn(I1) binding could then be evaluated from the disappearance of the free Mn(H20)6’+ or the appearance of the L-site spectrum. The former was standardized against MnCl2 of known concentration and provided a measure of the total Mn(I1) bound, i.e., both to the high affinity L sites and the low affinity N sites (Bagshaw and Reed, 1976). Plf =

T A B L E I: Relative D T N B Light Chain Content and L-Site

Stoichiometry of Native and DTNB-Treated Myosins. Myosin Sample”

Nativeb

DTNBEDTA

DTNBCa(I1)

Reconstituted

D T N B light chain L site

l.OOr l.OOd

0.59 0.61

1 .oo 0.95

0.92 0.89

“See Materials and Methods section for the terminology of myosin samples. bThe data are normalized with respect to native myosin. CCorresponds to 2 mol of D T N B light chain per myosin molecule (Weeds and Lowey, 1971). dCorresponds to 1.8 mol of Mn(I1) bound with high affinity (Figure 1).

Results

Mn(ZI) Binding to the D T N B Light Chain. In a previous study it was noted that the stoichiometry of Mn(I1) binding to the high affinity site of subfragment 1 depended on the conditions used for its preparation. Inclusion of a divalent metal ion (Mg(II), Ca(II), or Mn(I1)) during the papain digestion of myosin increased the fraction of subfragment 1 molecules bearing an intact DTNB light chain and high affinity Mn(I1) binding site, compared with subfragment 1 preparations obtained by digestion in the presence of EDTA (Bagshaw and Reed, 1976). While no major difference was detected in the heavy chain content of these preparations by sodium dodecyl sulfate gel electrophoresis (see also Margossian et al., 1975), the possibility remained that the Mn(I1) binding site was located on the heavy chain and its correlation with the DTNB light chain was coincidental. It is therefore important to demonstrate this relation with myosin preparations which have not undergone proteolysis and. if possible, with the isolated D T N B light chain. Selective removal of the DTNB light chain from myosin was achieved by DTNB-EDTA treatment as described in the Materials and Methods section. The myosin samples were subject to sodium dodecyl sulfate gel electrophoresis and their light chain contents were determined by densitometry (Table I). Mn(I1) binding to these samples, as determined by EPR spectroscopy, is illustrated in the Scatchard plots of Figure 1. Native myosin binds 1.8 mol of Mn(I1) with high affinity (KL < 1 pM). The equilibrium constant is a n upper limit owing to the competition from contaminating Ca(I1) and Mg(I1). Atomic absorption measurements indicate the metal binding site is initially near saturated with Ca(I1). DTNB-EDTA treatment reduces the stoichiometry of Mn(I1) binding to 1.1 mol, but a stoichiometry of 1.6 mol is restored if the sample is treated with dithiothreitol prior to precipitation of the myosin (reconstituted myosin). Mn(I1) binding measured by the appearance of the L-site spectrum (open symbols) parallels the Scatchard plots for Mn(I1) binding to the high affinity sites (solid symbols). For a system with one class of sites, the superposition of the solid and open symbols is expected since this factor depends only on the accuracy of measuring the peak heights (